Non-enhanced CT Dataset for Pediatric Extracranial Germ Cell Tumor Subtype Classification

A public dataset of 642 non-enhanced CT scans covering six pediatric extracranial germ cell tumor subtypes is now available, with preliminary radiomics models showing utility for AI-based diagnosis.

• Dataset includes expert-annotated tumor masks and standardized preprocessing for six histological subtypes: mature teratoma, immature teratoma, yolk sac tumor, mixed germ cell tumor, dysgerminoma, and embryonal carcinoma.
• Preliminary radiomics-based machine learning experiments demonstrate feasibility for developing subtype classification tools.

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.